D E C E M B E R 2 0 1 4 | w w w . t m f o r u m . o r g

E X T R A I N S I G H T S

NFV: WHAT DOES IT TAKE TO BE AGILE? DEVOPS TRANSFORMATION FRAMEWORK FOR THE DIGITAL ECOSYSTEM

N F V : W H A T D O E S I T T A K E T O B E A G I L E ?

2 w w w . t m f o r u m . o r g

N F V: W H AT D O E S I T TA K E TO B E AG I L E ? DEVOPS TRANSFORMATION FRAMEWORK FOR THE DIGITAL ECOSYSTEM

CONTENTS

3 INTRODUCTION

5 CHAPTER 1: Why is agility imperative for network operations?

8 CHAPTER 2: What are the lessons learned from IT operations?

10 CHAPTER 3: What does it take to go from concept to operations?

13 CHAPTER 4: How can your company adapt DevOps methodology for network operations?

17 CHAPTER 5: What is the organizational impact of adopting DevOps methodology?

21 CHAPTER 6: What should you do next?

N F V : W H A T D O E S I T T A K E T O B E A G I L E ?

3 w w w . t m f o r u m . o r g

In a traditional IT environment, applications development

teams go through a lengthy process of gathering

requirements followed by coding and testing before

an app can go into production. Not only is this process

time-consuming, it often takes place with little if any

consideration of the operating environment in which the

application will run. The introduction of DevOps makes this

process more agile and less time-consuming.

Wikipedia defines DevOps as “a concept dealing

with, among other things: software development,

operations, and services. It emphasizes communication,

collaboration, and integration between software

developers and information technology (IT) operations

personnel. DevOps is a response to the interdependence

of software development and IT operations. It aims to

help an organization rapidly produce software products

and services.” Since the term can be interpreted in many

ways, to avoid confusion this Extra Insights publication

accepts this definition.

As the name implies, DevOps is a combination

Introductionof development and operations: The applications

development team continuously writes small incremental

pieces of code that are continuously tested on an

architecture that reflects the production architecture.

Because IT organizations have had so much success with

this model, many executives are now looking at how to

apply lessons learned to network operations.

In this primer TM Forum introduces the concept of a

DevOps transformation framework, and in doing so

sets a clear vision for what the digital world requires to

operate networks made up of a combination of current

and virtualized infrastructure. Specifically, the framework

outlines how to transform from the current network

operations model to an enhanced one that drastically

reduces the time and resources it takes to introduce

new services and improves a service provider’s ability to

compete in the digital world.

DevOps transformation framework

N F V : W H A T D O E S I T T A K E T O B E A G I L E ?

4 w w w . t m f o r u m . o r g

I N T R O D U C T I O N

We will answer the following questions about moving

toward a DevOps model for network operations:

1. Why is agility an imperative for network operations?

2. What are the lessons learned from IT operations?

3. What does it take to go from concept to operations?

4. How can your company adapt DevOps methodology

for network operations?

5. What is the organizational impact of adopting a

DevOps approach?

6. What should you do next?

To put this publication into context, let’s consider an

important ‘user story’:

As a leader in the CTO/CIO organization accountable

for transforming my company so that it can take

advantage of virtualization, I need to move to an

operations approach that supports a dramatically more

agile service lifecycle, one that lowers capital and

operating expenses and supports continuous service

innovation and delivery.

To do this, I need to understand how operations must

change so that I can manage hybrid environments.

TM Forum’s ZOOM program is advancing a large body

of work underpinning four key themes for virtualization,

which include: moving away from the traditional service

provider operations model to a more agile DevOps model;

focusing on end-to-end virtual network and operations

management; working on NFV readiness support for

procurement and operations; and clarifying how open

source technology can be used. Much of the team’s work

is delivered as part of the Frameworx suite of tools and

best practices.

The team has delivered an assessment of how

virtualization impacts service level agreements and is

currently working on information, data and policy models;

NFV preparedness; Catalyst projects solving real-world

implementation issues; an end-to-end security fabric

for NFV; and a set of operational support system (OSS)

design principles needed for NFV adoption to become

widespread. The team calls this work ‘OSS Futures’, and

it draws heavily on existing TM Forum assets including

the Digital Service Reference Architecture, the Software-

enabled Services Management Solution and the B2B2X

Partnering Accelerator. For more information, please

contact Dave Milham via dmilham@tmforum.org or

Ken Dilbeck via kdilbeck@tmforum.org.

N F V : W H A T D O E S I T T A K E T O B E A G I L E ?

5 w w w . t m f o r u m . o r g

Virtualization of compute, storage and network resources

has many significant benefits for communications service

providers: It lowers capital and operating costs and helps

them develop new services more quickly; provides just-in-

time capacity to meet dynamic changes in user demand

for services; and delivers a higher-quality experience for

the end customer.

As service providers begin to transform their networks

to embrace virtualization in the form of software-defined

networking (SDN) and network functions virtualization

(NFV), their network operations teams need to adopt IT’s

DevOps practices to maximize the potential benefits of

the evolving software-defined infrastructure. This includes

adopting new tools that enhance agility, implementation of

agile operational and organizational models and procedures,

and in some cases adopting a new culture.

Why is agility imperative for network operations?

C H A P T E R 1

N F V : W H A T D O E S I T T A K E T O B E A G I L E ?

6 w w w . t m f o r u m . o r g

In a traditional environment, implementing services

such as network optimization or security is cumbersome

because it requires acquiring network appliances and

cabling them together in the correct order. Since each

appliance has its own unique interface, configuration is a

time-consuming, error-prone task.

As we explain in the second primer in this series,

C H A P T E R 1 – W H Y I S A G I L I T Y I M P E R A T I V E F O R N E T W O R K O P E R A T I O N S ?

“DevOps has always been about culture, not just about tools and processes. We found that the cultural practices and norms that characterize high-trust organizations — good information flow, cross-functional collaboration, shared responsibilities, learning from failures and encouragement of new ideas — are the same as those at the heart of DevOps.”Puppet Labs’ 2014 State of DevOps Report

NFV: Can it be managed? The blueprint for end-to-end

management, the deployment of virtualized network

functions (VNFs) improves this process significantly.

For example, it is now possible to dynamically create

optimization or security as a virtual function that can

be moved around from server to server and rapidly

decommissioned when it’s no longer needed.

“Network operations teams need to adopt IT’s DevOps practices to maximize the potential benefits of the evolving software-defined infrastructure.”

N F V : W H A T D O E S I T T A K E T O B E A G I L E ?

7 w w w . t m f o r u m . o r g

C H A P T E R 1 – W H Y I S A G I L I T Y I M P E R A T I V E F O R N E T W O R K O P E R A T I O N S ?

The adoption of VNFs has a dramatic impact on end-to-

end management (again, see NFV: Can it be managed? )

and requires many changes. The new tools, models,

procedures and culture adopted by the network operations

team must be able to respond to them. For example:

n End-to-end management systems must be re-

architected to support dynamic creation and

decommissioning of network functions in accordance

with a set of highly granular policies designed to ensure

quality of experience for the end customer.

n The fact that VNFs can be dynamically created and

moved makes it difficult to identify where a VNF is

running.

n In contrast to being negotiated and included in business

agreements for network services, service level

agreements (SLAs) are likely to be negotiated on the fly

as VNFs are chained together or as configurations are

modified.

n Topology changes, which occur infrequently in a

traditional environment, occur frequently in a virtualized

environment.

n Setting quality of service goes from being a laborious,

one-time manual process to a dynamic process that is

repeated as needed.

n The focus on fault management is replaced by a focus on

network and application performance.

n In a traditional environment, end-to-end management

has to be performed across technology domains. In the

emerging environment, end-to-end management happens

across technology domains that span both current and

virtualized infrastructures and in some instances multiple

service providers’ domains.

In the next chapters, we’ll provide guidance on how

DevOps processes can help service providers address many

of these necessary changes.

Time for change

“Adoption of virtual network functions has a dramatic impact on end-to-end management and requires many changes.”

N F V : W H A T D O E S I T T A K E T O B E A G I L E ?

8 w w w . t m f o r u m . o r g

One of the problems in traditional software development

is how to address the conflicting goals of rapidly

deploying new capabilities versus maintaining a highly

stable, error-free environment. IT operations teams that

have successfully adopted DevOps methodology have

managed to overcome this obstacle. When asked in a

2014 InformationWeek survey to indicate the level of

improvement in application development speed that they

had either already gained or expected to gain as a result

of adopting DevOps, 41 percent of respondents indicated

‘significant improvement’ and 42 percent noted ‘some

improvement’.

To apply DevOps to network operations, it’s important

to understand the principles used in a successful DevOps

software development environment:

What are the lessons learned from IT operations?

C H A P T E R 2

n Collaboration – a key aspect of DevOps requires

creating a culture of collaboration among all the groups

that have a stake in the delivery of new software.

n Continuous integration and delivery – software

changes are added to a large base of code immediately

after development so that new capabilities can

be continuously delivered to the entire chain of

users for testing and monitoring in production-style

environments. The goal is to generate rapid feedback so

that defects can be identified and corrected as soon as

possible.

n Continuous testing and monitoring – testing is

performed continuously at all stages of the release

process, and not just by the quality assurance team.

Developers perform testing and provide test data and

procedures that can be used by all collaborating groups.

N F V : W H A T D O E S I T T A K E T O B E A G I L E ?

9 w w w . t m f o r u m . o r g

C H A P T E R 2 – W H A T A R E T H E L E S S O N S L E A R N E D F R O M I T O P E R A T I O N S ?

The operations group is also typically involved in testing

and monitoring. This team can add value by specifying

load patterns to make test environments more realistic,

and they can monitor continuously to identify service

issues and resolve them in near real time. Note that

the same set of tools that monitor the production

environment can also be used in development to identify

potential performance issues.

n Automation – all stages of software delivery are

highly dependent on automated tools. Automation is

essential because it enhances agility and provides the

productivity required to support the continuous nature

of integration, delivery, testing and monitoring of many

small incremental changes to the code base.

n Automated management interfaces – software-defined

environments (SDEs) are an emerging core capability of

DevOps that allow organizations to manage the scale and

speed with which environments need to be provisioned

and configured to enable continuous delivery. SDEs use

technologies such as automated management interfaces

in the form of application program interfaces (APIs) that

define entire systems made up of multiple components.

These interfaces are based on information models that

define the characteristics, behaviors, configurations,

roles, relationships, workloads and workload policies for

all the entities that comprise the system.

In the IT world, applying DevOps principles has delivered

significant benefits. It provides enhanced customer

experience for service differentiation, which is maximized

by incorporating customer feedback as an important

aspect of the DevOps continuous monitoring function.

Organizations also enjoy increased capacity for innovation

because they can leverage emerging technologies, such as

cloud computing, mobile applications, big data and social

media. Finally, they realize faster time to value because they

have adopted the methodology, tools and culture required

to minimize the time between conception and delivery.

TM Forum does a significant amount of work around APIs. For more information check out the API Zone. We’ll discuss information

models more in the next chapter, but you can find additional information in the second primer in this series, NFV: Can it be managed?

and in ZOOM’s Why information models are needed for agile operations.

N F V : W H A T D O E S I T T A K E T O B E A G I L E ?

1 0 w w w . t m f o r u m . o r g

What does it take to go from concept to operations?

C H A P T E R 3

In order for service providers to realize the value of virtual

and hybrid networks, operations staff must develop

additional skill sets and adopt more agile operational and

organizational models. Managing virtual network functions

running on virtual machines requires skills and processes

similar to IT’s best practices for data center operations

– for example, automatically spinning up a VM when

utilization crosses a pre-set threshold.

To maximize the benefits of policy-driven automation

in network operations, service providers must be able to

dramatically reduce the amount of time it takes to develop

and introduce new network services supported by a range

of VNFs from a variety of suppliers. As the adoption of

virtualization progresses, end-to-end management systems

and other OSS/BSS capabilities will have to evolve rapidly

to keep up.

N F V : W H A T D O E S I T T A K E T O B E A G I L E ?

1 1 w w w . t m f o r u m . o r g

C H A P T E R 3 – W H A T D O E S I T TA K E T O G O F R O M C O N C E P T T O O P E R A T I O N S ?

“Our experience on this journey to date has been that the small, self-directed teams required in a DevOps world require an amalgamation of skills spanning everything from IT security to database design and application architecture, plus everything in between. While each individual on the team has a particular strength (say, application design and coding), each one also needs to have working knowledge in other areas (maybe UX or network design).”Eric Reed, CTO, GE Capital as quoted in InformationWeek

The use of DevOps in IT operations has taught us

that continuous integration and testing is required and

that these processes must be automated. To achieve

automation, an information model is required to develop

the necessary management interfaces.

An information model is a way of describing a wide range

of entities, such as network functions (for example, firewall

or routing functions), customers and product information.

The description includes the characteristics and properties

of each of the entities and their relationships with others.

A data model is a specific representation of an information

model. For example, a data model may describe the exact

schema used to represent customer information in a

database or to exchange customer information across an

interface between two or more OSSs.

When models are implemented correctly, it’s possible

to precisely combine information from multiple suppliers

and multiple technologies to automate service operations

and create consistent end-to-end views. Very often this

doesn’t happen, however, which leads to the need for

labor-intensive OSS integration. This is a major reason why

existing OSS/BSS solutions are inflexible and expensive.

Continuous integration and testing

N F V : W H A T D O E S I T T A K E T O B E A G I L E ?

1 2 w w w . t m f o r u m . o r g

C H A P T E R 3 – W H A T D O E S I T TA K E T O G O F R O M C O N C E P T T O O P E R A T I O N S ?

In order to implement a DevOps-like continuous

development model, service providers must determine how

they will ensure interoperability. Standards traditionally have

been defined by creating a detailed set of requirements and

a single, specific implementation with supporting test tools.

This assumes that all the requirements are known at the

outset, which is rarely the case.

Using core common information models that incorporate

best practices is a more flexible way of ensuring

interoperability. TM Forum’s Information Framework,

which is part of the Frameworx suite of standards-based

tools and best practices, is an example. The ZOOM team

is working to enhance this model for virtualized and hybrid

networks by incorporating lessons learned from several

Catalyst proof-of-concept projects.

The Service bundling in a B2B2X marketplace Catalyst,

which was demonstrated at TM Forum Live! in June,

How to ensure interoperability?showed how a buyer can bundle a collection of services

sourced from different suppliers and deliver them seamlessly

to a customer in a business-to-business or business-to-

business-to-consumer arrangement.

A second TM Forum Live! Catalyst called CloudNFVTM:

Dynamic, data-driven management and operations,

delivered a metamodel for event-driven management and

operations, and included a live demonstration of a dynamic

implementation of a network service as a VNF deployed via

OpenStack cloud orchestration and optimized to maintain

quality of service. That Catalyst is continuing at Digital

Disruption under a new name, Preparing NFV for prime time.

A key result of the CloudNFVTM project was a contribution

to the Information Framework on metamodel extensions to

support dynamically defined management APIs, while the

B2B2X project contributed a similar metamodel extension

along with a set of guidelines and examples.

TM Forum Catalysts are short-term, collaborative projects led by our members that create solutions for addressing today’s

most pressing operational and systems challenges. For more about the Catalysts mentioned here, see the second primer in

this series NFV: Can it be managed? or check out this article on Inform.

N F V : W H A T D O E S I T T A K E T O B E A G I L E ?

1 3 w w w . t m f o r u m . o r g

Current network service development and operations are

based largely on manual and/or static techniques, such as

those required to configure quality of service parameters

on switches and routers. However, the transition to

virtualized network infrastructure and network services

based on VNFs creates a more flexible environment similar

to DevOps-based IT operations.

In a virtualized environment, network services are

created by linking together, or chaining, VNFs that may

have been sourced from a variety of suppliers. One of the

primary goals of NFV is to streamline the integration of

these VNFs. For example, a service provider may purchase

network optimization functionality from one supplier and

security from another.

It is important to minimize or eliminate the need for

manual intervention of any sort when incorporating VNFs.

The Service bundling in a B2B2X marketplace highlighted

in Chapter 3 (see page 12) demonstrated how to bundle a

How can your company adapt DevOps methodology for network operations?

C H A P T E R 4

N F V : W H A T D O E S I T T A K E T O B E A G I L E ?

1 4 w w w . t m f o r u m . o r g

C H A P T E R 4 – H O W C A N Y O U R C O M P A N Y A D A P T D E V O P S M E T H O D O L O G Y F O R N E T W O R K O P E R A T I O N S ?

collection of services from different suppliers. In this case

shared product definitions were used to combine a cloud

service, a software-defined network and firewall mocked

up by Cisco Systems with fiber access service provided

by NBN Co of Australia. A catalog management system

provided by DGIT dynamically configured these three

components into a managed bundled service with catalog-

defined, dynamically rendered management APIs.

All of the basic DevOps principles discussed in Chapter 2 are

applicable in a NetOps setting. However, DevOps is generally

applied to discrete services that are frequently delivered over

the Internet on a best-effort basis. Communications service

providers, on the other hand, need to be able to ensure

quality of service and guarantee service level agreements

(SLAs) end to end – often across multiple partners’ networks.

Another Catalyst demonstrated at TM Forum Live! in

June called Orchestrating SDN and NFV while enforcing

an SLA over a WAN highlighted some of the limitations

of applying DevOps to network operations. The Catalyst

used a DevOps ‘service mesh’ and use cases provided

by the Open Data Center Alliance to demonstrate SLA

management in a cloud environment.

The Catalyst highlighted the difficulty of ensuring

compatibility between VNFs. When VNFs are

chained together for an end-to-end service, strong

interdependencies between them are created. For

example, if a service provider updates an optimization VNF,

it needs to ensure that it is fully compatible with other

VNFs, such as those used for security. As a result, much

stronger version control and compatibility testing is needed

than would be typical in an enterprise application.

The Catalyst is continuing at Digital Disruption with

Microsoft and AT&T championing it in the hopes that they

can learn how to eliminate the necessary upfront planning

now required to ensure SLAs in a multi-cloud environment.

For more about the Catalyst, see the second primer in this

series NFV: Can it be managed? and this article on Inform.

Ensuring quality end to end

N F V : W H A T D O E S I T T A K E T O B E A G I L E ?

1 5 w w w . t m f o r u m . o r g

C H A P T E R 4 – H O W C A N Y O U R C O M P A N Y A D A P T D E V O P S M E T H O D O L O G Y F O R N E T W O R K O P E R A T I O N S ?

Other challenges in the NetOps world that need to be

addressed by a new hybrid DevOps model include:

n Since the vast majority of environments will be

hardware- and software-based for the foreseeable

future, NetOps methodology must accommodate

services that depend on network functions running on

dedicated hardware platforms as well as VNFs.

n Virtualized services will often be created by integrating

services from multiple suppliers. This will require

NetOps and DevOps methodologies and best practices

to support concurrent, synchronized development and

integration across the domains of multiple partners

using dynamic APIs.

n In contrast to the Internet’s best-effort service delivery,

NetOps will need to support dynamic and automated

management of service performance and SLAs. This

can only be achieved by a policy model that supports

end-to-end SLA targets.

n Again, in contrast to what happens when delivering an

application over the Internet, NFV services are often

mission critical. This creates a need for high levels

of resilience and rapid fallback capabilities to restore

services.

n Virtualized services will cover a wide range of network

functions and technologies. As a result, consistent

frameworks and interfaces are needed across suppliers’

solutions to minimize or eliminate the need for manual

intervention of any sort when incorporating VNFs into a

network service.

n Because of the large number of network functions and

technologies, creating NetOps test environments that

accurately simulate production environments is much

more difficult to do than in a DevOps environment.

It’s not going to be easy

“NetOps will need to support dynamic and automated management of service performance and SLAs.”

N F V : W H A T D O E S I T T A K E T O B E A G I L E ?

1 6 w w w . t m f o r u m . o r g

C H A P T E R 4 – H O W C A N Y O U R C O M P A N Y A D A P T D E V O P S M E T H O D O L O G Y F O R N E T W O R K O P E R A T I O N S ?

ZOOM’s vision for NFV is based on the ability to flexibly

integrate a wide variety of suppliers’ solutions covering

multiple technologies. Several other standards-defining

organizations (SDOs), collaboration groups and open

source groups are tackling the challenges as well, but

inevitably the work is diverging, which means that multi-

technology integration is becoming more complicated. It

is clear that some umbrella organization needs to drive the

multi-technology, converged management agenda.

TM Forum is well positioned to help in this effort.

The Forum’s OSS/BSS futures framework leverages

best practices and experience from many Catalyst

demonstrations featuring collaboration between the Forum

and groups such as the 3rd Generation Partnership Project,

the Alliance for Telecommunications Industry Solutions,

the European Telecommunications Standards Institute,

the Internet Engineering Task Force, the Next Generation

Mobile Network Alliance and the Open Networking

Foundation.

In addition, the Forum’s Digital Services Reference

Architecture (DSRA) proposes a simple set of modeling

principles and API patterns, which if widely adopted

would alleviate the current integration challenge and lay

the foundation for automated APIs that can be integrated

by configuration rather than coding to support DevOps

for network operations. Such a capability is an essential

pre-requisite for achieving automated development and

integration, and without widespread adoption of these

principles, solutions will be unable to support the necessary

levels of automation to deliver on the promise of NFV.

ZOOM can help

“ It is clear that some umbrella organization needs to drive the multi-technology, converged management agenda.”

N F V : W H A T D O E S I T T A K E T O B E A G I L E ?

1 7 w w w . t m f o r u m . o r g

The greatest organizational challenge in applying DevOps

methodology lies in fostering effective collaboration across

all the groups involved in bringing network services to

market, including standards-defining organizations.

There are a number of possible approaches to creating

a culture of closer collaboration. One that has been used

successfully within IT DevOps is merging some or all

of the application developers with some or all of the

operations organization to create a single, multi-functional

team with a set of performance incentives and metrics

that align closely with the overall business objectives

of the organization – for example, reduced time to

deployment for new applications.

An alternative approach is to leave the basic

organizational structure in place and focus on replacing

conflicting objectives and incentives with a common set of

metrics, such as for measuring the quality of applications.

In either case, collaboration is greatly enhanced by a

common set of automated tools that span organizational

and/or technological boundaries. With both types of

organizational structure, new processes must be developed

to support the DevOps practices of continuous integration,

delivery, testing, monitoring and feedback. Where these

new practices are based on well-defined and automated

processes, business objectives for agile development of

innovative, high-quality services can be met more efficiently.

What is the organizational impact of adopting DevOps methodology?

C H A P T E R 5

N F V : W H A T D O E S I T T A K E T O B E A G I L E ?

1 8 w w w . t m f o r u m . o r g

C H A P T E R 5 – W H A T I S T H E O R G A N I Z A T I O N A L I M P A C T O F A D O P T I N G D E V O P S M E T H O D O L O G Y ?

In a recent survey, 33 communications service provider

executives were asked about the impact NFV will have

on the organization of their companies over the next two

years. More than 44 percent indicated it would have either

a ‘significant’ or ‘very significant’ impact.

Some of the answers from service provider respondents

when asked to indicate the type of organizational changes that

have already occurred or that they expect to occur include:

n The operations group is likely to be restructured;

n The company’s technical experts have been

consolidated into a single group;

n The company has set up a subsidiary and is in the

process of moving IT employees to that subsidiary; and

n The organization’s OSS/BSSs need to be revamped.

“ In the forward-looking CTO group where we’re looking at technology down the road, we can’t make decisions or requirements in a vacuum – we have to be able to have everybody’s review of it. We need to eliminate silos between organizations. We have included the IT group in our proof of concepts and in writing framework documents for deployment, and we also have the network team involved in helping review test plans. We’re getting input from everyone on RFIs, and we’re learning how to work together across the borders.”C-level Executive, Tier 1 Service Provider

Change is underway HOW WILL NFV IMPACT ORGANIZATIONS OVERALL?

Source: Ashton, Metzler & Associates

VERY SIGNIFICANT IMPACT (14.81%)

SIGNIFICANT IMPACT (29.63%)

NO IMPACT(7.41%)

MODERATEIMPACT

(33.33%)

SOME IMPACT(14.81%)

DON’T KNOW (0%)OTHER (0%)

N F V : W H A T D O E S I T T A K E T O B E A G I L E ?

1 9 w w w . t m f o r u m . o r g

C H A P T E R 5 – W H A T I S T H E O R G A N I Z A T I O N A L I M P A C T O F A D O P T I N G D E V O P S M E T H O D O L O G Y ?

Survey respondents were also asked how much of an

impact they think NFV will have on the nature of their jobs

over the next two years. Nearly 60 percent said it would

have either a ‘significant’ or ‘very significant’ impact.

Some of the answers from service provider respondents

when asked to indicate the type of changes that have already

occurred or that they expect to occur to their jobs include:

n The product development lifecycle will change;

n The job will require new skills in general and more

knowledge of software in particular;

n There will likely be a new, unknown set of customer

demands;

n Product development needs to be able to provide tools

to manage and monitor the environment; and

n There will be new business models and new product

offerings that must be supported.

HOW WILL NFV IMPACT JOBS?

Source: Ashton, Metzler & Associates

VERY SIGNIFICANT IMPACT (22.22%)

DON’T KNOW (3.70%)

SIGNIFICANT IMPACT (37.04%)

MODERATEIMPACT

(18.52%)

SOMEIMPACT

(11.11%)

OTHER (3.70%)

NO IMPACT (3.70%)

“Jobs will require new skills and more knowledge of software, and there will be a new, unknown set of customer demands.”

N F V : W H A T D O E S I T T A K E T O B E A G I L E ?

2 0 w w w . t m f o r u m . o r g

C H A P T E R 5 – W H A T I S T H E O R G A N I Z A T I O N A L I M P A C T O F A D O P T I N G D E V O P S M E T H O D O L O G Y ?

The network operations team’s use and enhancement of

DevOps practices and automated processes will impact

the activities and organizational structure of the entire

service-delivery chain. For example, changing the focus

of service-development activities from code generation

to VNF chaining via configuration will require significant

structural changes for existing development groups. In

addition, many network organizations are adopting the

DevOps practice of removing middle line managers in

favor of setting up self-organizing teams with end-to-end

responsibility for a specific feature.

Another significant organizational challenge is the need

to support the rollout of network services that partially

depend on VNFs hosted by partners. This requires

organizations to be able to integrate and synchronize

NetOps processes across the boundaries between

administrative domains. For this to be feasible, new

standards and best practices will be needed to share

responsibilities and exchange data between the automated

systems of multiple partners.

Since current DevOps practices do not address

enterprise development synchronization, there clearly is

a need for best practices and standards for the exchange

of product and service definitions, lifecycle state and

other development and operations metadata. The Forum’s

work on the Digital Services Reference Architecture, the

Software-enabled Services Management Solution and

the B2B2X Partnering Accelerator provide a starting point

for addressing these challenges, and potential solutions

using these best practices have been demonstrated in the

Catalysts discussed in Chapter 3 (see page 12).

The ZOOM team is incorporating the lessons learned

from the Catalysts in the next iterations of Frameworx,

which will help service providers make a successful

transition to DevOps-based network operations.

Addressing the challenges

“Best practices will be needed to share responsibilities between automated systems of multiple partners.”

N F V : W H A T D O E S I T T A K E T O B E A G I L E ?

2 1 w w w . t m f o r u m . o r g

Hopefully this primer has helped you understand the

opportunities and challenges associated with adopting a

modified DevOps approach to network operations. The

benefits of adopting such an approach are compelling:

lower costs and increased agility, including the ability

to implement new services in days rather than weeks

or months. Unfortunately, the challenges are significant

and include extending existing information models,

adopting new tools, implementing new operational and

organizational models, and in some instances, changing a

company’s culture.

TM Forum’s ZOOM team has done a significant amount

of work already, creating best practices and experiences,

crafting blueprints for operational models, extending the

information model, modifying the DSRA and demonstrating

practical solutions through the Catalyst program.

A lot of hard work remains, but in the end it will be worth

the effort, because it will help you reduce costs and serve

your customers better and more quickly, with the kinds of

services they are demanding.

The next steps to take include learning more and getting

involved. To learn more about NFV readiness and the

necessary changes to end-to-end management, check out

the first and second eBooks in this series. You can find

them here. You can also find more detailed information

about DevOps and more information about ZOOM here.

Finally, get involved! Don’t be content to wait for

virtualization technology to mature – join the ZOOM team

and make your voice heard as we set the course for the

future of networking by developing the tools and best

practices required to make network virtualization a success.

What should you do next?

C H A P T E R 6

T M F O R U M

2 2 w w w . t m f o r u m . o r g

© 2014. The entire contents of this publication are protected by copyright. All rights reserved. The Forum would like to thank the sponsors and advertisers who have enabled the publication of this fully independently researched report. The views and opinions expressed by individual authors and contributors in this publication are provided in the writers’ personal capacities and are their sole responsibility. Their publication does not imply that they represent the views or opinions of TM Forum and must neither be regarded as constituting advice on any matter whatsoever, nor be interpreted as such. The reproduction of advertisements and sponsored features in this publication does not in any way imply endorsement by TM Forum of products or services referred to therein.

REPORT AUTHOR:Jim Metzler, Ph. D.Founder and Vice President, Ashton, Metzler & Associatesjmetzler@ashtonmetzler.com

SENIOR DIRECTOR, CONTENT:Annie Turneraturner@tmforum.org

EDITORS:Dawn Bushausdbushaus@tmforum.org

Lisa Hugheslhughes@tmforum.org

BUSINESS DEVELOPMENT DIRECTOR,RESEARCH & PUBLICATIONS:Mark Bradburymbradbury@tmforum.org

PUBLISHER:Amy Gogginsagoggins@tmforum.org

SENIOR MARKETING MANAGER, SOLUTIONS MARKETING: Charlotte Lewisclewis@tmforum.org

MANAGER, DIGITAL CONTENT: Sarah Wrayswray@tmforum.org

REPORT DESIGN:thePAGEDESIGN

ADVISORS:Nik Willetts, Chief Digital Officer, TM ForumRob Rich, Managing Director, Insights Research, TM ForumCarl Piva, Vice President, Strategic Programs, TM ForumDave Milham, Chief Architect, Service Provider Engagement, TM Forum

PUBLISHED BY:TM Forum240 Headquarters PlazaEast Tower, 10th FloorMorristown, NJ 07960-6628USA

www.tmforum.org

Phone: +1 973-944-5100Fax: +1 973-944-5110

ISBN: 978-1-939303-68-4

E X T R A I N S I G H T S